Knowledge Representation

Superordinate (global) Level

"furniture"

Subordinate (specific) Level

"kitchen table"

Basic Level**

"table" -going above to superordinate results in LARGE LOSS of info -going below to subordinate results in LITTLE GAIN of info FASTER RESPONSE to basic level, depending on expertise and familiarity

Collin's and Quillian's Hierarchical Model

-semantic network model -nodes, links connecting concepts The time it takes for a person to retrieve info about a concept should b determined by the distance that must be traveled though the network -FASTER for "a canary is a bird" only have to travel one link -SLOWER for "a canary is an animal" bc have to travel two links

Connectionism

1. Based on how info is represented in the brain 2. Can explain a # of findings, including how concepts are learned and who damage to the brain affects ppl's knowledge about concepts Approach to creating computer models for representing cognitive processes

Criticism of Collins and Quillian Model

1. Can't explain typicality effect and reaction times for more typical members and less typical members 2. Cognitive economy 3. Traveling links and response time

High Typicality: Rosch's point scale

A category member closely resembles that category prototype -sparrow

Typicality Effect

Ability to judge highly prototypical objects more rapidly

Spreading Activation

Activity that spreads out along any link that is connected to an activated node -the additional concepts that receive this activation become PRIMED and so can be retrieved more easily from memory

Exemplar -later in learning -small categories (U.S. presidents) -use real examples-->can take into account more information

Actual member of the category that a person has encountered in the past -if experienced sparrows, robins, and blue jays in past, each would be an exemplar for the category "birds"

Low Typicality

Category member does NOT closely resemble a typical member of the category -bat, penguin

Activation of Units in a Network

Depend on: 1. the signal that originates in the input units 2. connection weights throughout the network

Connection Weight

Determines how signals sent from one unit either increase or decrease the activity of the next unit High connection weight: strong tendency to excite the next unit Low connection weight: cause less excitation Negative weight: can decrease excitation or inhibit activation

Definitional Approach to Categorization

Determining whether a particular object meets the definition of the category -"square" works well -chairs, birds, tress, do NOT work well --not all members of same category have same features

Graceful Degradation

Disruption of performance occurs only gradually as parts of the system are damaged

EXPERTS: Knowledge can affect Categorization

Experts usually use subordinate level instead of basic

Prototypical Objects have HIGH Family Resemblance: Rosch and Mervis (1975)

FURNITURE -chair/sofa-->lots of overlap with characteristics of many items in a category, family resemblance HIGH -mirror/telephone-->little overlap

Naming

High-prototypical items are named first -Sparrow before penguin

Family Resemblance

Idea that things in particular category resemble one another in a number of ways -allows for some variation -determine how similar an object is to some standard representation of a category

Exemplar Approach to Categorization

Involves determining whether an object is similar to other objects Involves MANY examples, each one called an exemplar NOT AVERAGE

Conceptual Knowledge

Knowledge that enables us to recognize objects and events and tomato inferences about their properties

Prototype Approach to Categorization

Membership in a category is determined by comparing the object to a prototype that represents the category -AVERAGE

Multiple-Factor Approach

Neural activation represents similarities and differences in the kinds of features/information contained with categories Idea of distributed representation Focuses on searching for more factors that determine how concepts are divided up within a category

Sensory-Functional (S-F) Hypothesis

Our ability to differentiate living things and artifacts depends on a semantic memory system that distinguishes sensory attributes and a system that distinguishes function -TOO SIMPLIFIED

Embodied Approach

Our knowledge of concepts is based on reactivation of sensory and motor processes that occur when we interact with the object

Category-Specific Memory Impairment

Patients lose the ability to identify one type of object but retained ability to identify other types of objects

Parallel Distributed Processing (PDP) Models

Propose that concepts are represented by activity that is distributed across a network 1) Based on functioning of nervous system 2) System not totally disrupted by damage (graceful degradation) 3) Learning can be generalized 4) Successful computer models have been developed Problems: 1) How does back-propogation work? 2) Learning one concept interferes with others 3) Rapid learning difficult to explain

Semantic Netowork Approach

Proposes that concepts are arranged in networks

Priming

Prototypical objects affected more by priming Hear the word "green" -shown circles -two circles primary green -two circles light green --quicker response to prototypical good color

Crowding

Refers to the fact that animals tend to share many properties (eyes, legs, ability to move) -artifacts like cars and boats share fewer properties

Cognitive Economy

Storing shared properties just once at a higher-level node -birds-->can fly and have feathers

Lexical Decision Task: Meyer and Schvanevedlt (1971)

Subjects read stimuli, some words some not words -indicate as quickly as possible if each entry is word or nonword -pairs of words closely associated (bread and wheat) -FASTER bc of spreading activation -some not closely associated

Hierarchical Organization

The kind of org. in which larger, more general categories are divided into smaller more specific categories, creating a number of levels of categories --basic level not same for everyone

Concepts

The mental representation of a class or individual AND the meaning of objects, events, and abstract ideas

Semantic Category Approach

There are specific neural circuits in the brain for some specific (evolutionarily important) categories

Sentence Verification Technique: Smith (1974)

To determine how rapidly ppl could answer questions about an object's category Given sentences to respond with yes/no -respond faster for objects HIGH in prototypically (apple) -slower for pomegranate

Input Units

Units activated by stimuli from ENVIRONMENT -send to hidden units-->send to output units

Prototype -early in learning -big categories (birds)

a "typical" member of a category -based on an average of members of a category that are commonly experienced -AVERAGE representation -birds